Shift November 08, 2019

AI The Game at Shift


Ward Van Driessche

Artificial Intelligence Engineer

How do you plan to apply artificial intelligence in your department next year? We kick-off most of our AI workshops with this seemingly simple question, which assumes that you should have a plan by now. We also asked this question at our last annual Shift boutique conference. The participants could join our machine learning team to to come up with innovative AI ideas by playing our card game ‘AI The Game’.

A group of 35 attendees with varying backgrounds were split into seven small groups, that each played one round of ‘AI The Game’. First, the smaller groups agreed on a company they would all be part of for the next hour. Then, each player came up with an AI use case for their imaginary or real company. To inspire each participants, each got a handful of cards from our game deck. These cards consist of a perspective (function) they play within this company, three machine learning technologies, and a stack of value cards that players can use to defend their idea.

They then had to pitch their idea to one another and funnel their best AI strategy into an idea that they want to use to pitch to the other teams. Our machine learning team was there to better explain several AI technologies to the teams and help guide the their brainstorm sessions into a company strategy.


These are the use cases the groups came up with

Colruyt - What will we eat today? (Winning Idea!) ‘Many of us struggle with the first world problem that we have a full fridge, but no idea what we can make for dinner with all that stuff. As, Colruyt, we want to be able to provide our customers with a suggestion of what they could make for dinner with whatever they have available. We could scan our fridge with our smartphone and use object detection to easily find recipes.’

Electrobel - Energy optimization and smart grid ‘This fictional company administers several energy sources distributed over Belgium. With this amount of data, creating an efficient energy grid is far from straightforward. With the power of machine learning, we might be able to predict when and where energy will be necessary. We could anticipate on this to consume energy that is (or will probably be) produced nearby, optimizing our grid.’

Unilin - Floor assistant

“We would like to build a virtual assistant that helps you pick the floor of your dreams. Just show us some Pinterest images, and we will match similar floors in our assortiment. The assistant will also create an AR estimation of how this floor will look in your house. If you have questions about which glue you should use, or how much, just ask your digital assistant.”

Colruyt - Client needs help!

‘Whoever had the feeling they were looking for a product, or needed somebody for guidance and there is no staff member in sight? We came up with an idea to use video analysis to find.’

Telenet - Multiple installation visits

“Telenet installers sometimes need to visit multiple times to complete an new installation. We try to predict the presence of the equipment that’s inside a house, so we can avoid multi-day installations. They are not fun for customers, and inefficient for Telenet.”

Volvo cars - Motivate

“Employee dissatisfaction is bad for everyone. Performance is low and the employee is not having a good time. Their idea here is to use employee activity data to measure their motivation. Adapt and deal with the situation before the employees will churn.“

Brussels airlines - Stress prediction

‘Airlines are busy places and need a lot of management. This stresses out many of our staff members but also our customers. Staff is in a constant rush of having to coordinate everything continuously with high risks when something goes wrong. We could develop a machine learning tool that, with the right data, is able to predict when a co-worker is feeling overloaded or when he or she is accumulating an unhealthy amount of stress. We could also use image recognition and pose estimation to try and pick out customers that are getting nervous rapidly and use this to try and calm them down before something escalates.’


What’s in it for you?

As you can see, AI The Game is a fast solution to generate AI use cases that can help you in your role, or lead your organization towards implementing AI.

Now ask yourselves: How are you going to apply AI in your department next year? If that’s hard to define, let’s get in touch and play AI The Game with us. Inspiration guaranteed!

Download our Shift report

The latest edition contains 10 articles on the next big trends in digital, from creative artificial intelligence and the AR cloud to antifragility and cross-platform development.